The temperature distribution in presence of gravity, as measured by a local observer, is given by the Tolman expression. Here I derive the same only from the Gibbon's-Hawking-York surface term. In this process no explicit use of Einstein's equations of motion is done. Therefore, the present one is an off-shell analysis. Finally I discuss the importance and various implications of the derivation.